Some insight into things learned on teardown and inspection on our Sube powered 6A. http://www.sdsefi.com/rv16.htm
Russ---are the cylinders tapered, with more skirt clearance near top?
Tom
I get the biggest kick out of tearing down an engine - so thanks for sharing!
Apparently no observable wear on cam or followers?
Put the old valve springs in?
I would bet the cyls are glazed. I have no idea of the phyiscs for that happening on a broke in, water cooled engine but I have had a couple of air cooled that had the exact same symptoms. Essentially no measurable wear on piston or cyls, rings that had lost temper, (obviously wirey instead of brittle), lots of blowby, oil that turned black more quickly that it should and high oil consumption. I can convince myself your cyls look just a little yellow.
Looks to me like the hone on the cylinders wasn't done very well. The angles look way too shallow. You might want to talk to your ring manufacturer and see what Ra (avg) Rpk (peak roughness) and Rvk (valley roughness) numbers to shoot for on the hone and check with a profilometer.
And are those push on fittings for the oil cooler??
When you think about the unusual rod bearing damage, consider how hard the crank is oscillating those rod throws during your 1100 to 1600 RPM resonant period.
The bearing damage would appear to be classic (but rare) cavitation damage. The other 2 bearings were so mint, they could have been reused.
+1 on bore polishing for your oil consumption increase. Heavy duty engines, will have an increase in top land deposits and with the temps and piston motion can polish the bores. It looks like it is mostly on the thrust sides of the bore.
Also, the higher temps for the top ring when running at a high load steady state (i.e. cruise) the ring gap may need to be increased. Suggest checking the gap for butting evidence under high magnification.
Was the rod bearing cavitation on all rods? Have you considered all aluminum bearings? They have much higher fatigue resistance than overlaid bearings, but require a very clean oil during early life, and after oil changes. Also, a one factor for cavitation is aeration of the lube oil. Is there any particular churning of the oil that may be occurring? Like crank dipping during climbs etc? You know the engine best.
Just some thoughts from my database.
Thanks for sharing your results in such detail, this could really help make some key modifications to make an non-traditional engine last longer.
So which ones were damaged? Close to the flywheel or most distant?
Did you happen to check the rig temper relative to new? I know it is hard too waste a set but it could be enlightening. If temper is gone, pretty clear that they get too hot in continuous high output ops. Maybe a different material would solve it or just do routine maintenance. Gaskets, rings and a hone every half a decade ain't so bad so long as it's predictable and non catastrophic.
Ceramic coatings on the piston tops might cut down on heat flux. Especially with unleaded fuel quick will leave fewer insulating deposits.
Personally I think mirror finish is a mistake, with the finish too smooth there is no oil retention.
Here is a 1989 paper on the 2.0, is it representative of the lube system?
http://www.surrealmirage.com/subaru/files/SAE_boxer.pdf
If so, are the cavitated rod bearings in #2-3 or #1-4 perhaps? #2-3 are fed from the center main bearing and #1-4 are fed from one main each. Cavitation typically occurs at a point on the bearing where the pressure drops following the combustion pressure point. A radial force diagram for the bearing relative to the rod would be the typical method of investigation. Most engines in the field never see this (cavitation) as it is cured in the development process.
Sooo, if it is on #2-3 rods, it could be improved with a combination of higher oil pressure, (relief valve spring) , stronger bearing material, or reducing aeration of the oil. The last would be the most difficult. Attention to oil passages may help too. You probably have already thought of these.
One last comment, racing engines, although high power density seldom run very long at peak power; turns, shifting etc affect their ability to reach peak stabilized temps, like juggling a hot potato and surface temps on ones hand. Even F1 are only about 50% load factor per lap. Oval tracks (Indy cars) are a different matter.
Good luck, really interesting stuff.
The 10w30 should improve the pressure at the bearing as well (over your 10w60). Good idea, you already are looking at temps too, so that should not be a factor.
That cost story sure balances the total Subbie equation much better that one (me) would think. Thanks for the quantification.
Last comment, i assume you checked for oil leakage from the turbo seals already.
Cheers!
One would think except the engine works fine for a very long time of hard use with almost zero oil consumption. In fact, most moly ring manufacturers recommend a minimum of 320 grit wall finish. Moly rings don't require as much lubrication as other ring types also. The 14 year Toyota 20R turbo engine had Corvair TRW forged pistons fitted with moly rings. Went through 3 friends owning it and maybe 5000 hours. At the end, was using 1/2 quart in about 5000km. Not bad.
Barb fittings, Hey it's a car engine! - yep, again never had one leak, fail or fracture in 35 years and a lot of race cars.
I know of an F1 Rocket that was totalled due to a barbed fitting coming apart. If for some reason your oil cooler plugged up the barbed fitting is guaranteed to fail. There are lots of manufacturers now making kevlar hoses that are very light, flexible, and can hold up to 2000psi. Why chance it?
I've talked to a number of ring manufacturers at the PRI show and they have really come around to specifying the cylinder wall finish in terms of measurable roughness. Not just in terms of average roughness but the roughness of the peaks and valleys. I can rough with a 150 grit diamond hone, for example, come back and hit the bore with a 600 grit stone and the Ra could work out somewhere to 320 but the Rvk and Rpk numbers are going to be way different. Thats why just giving a grit # doesn't really work. Differences in surface hardness are going to affect Rvk and Rpk even though you might be using a 320 grit aluminum oxide stone. Then the stones aren't equal either... 320 grit diamond stones and a 320 grit aluminum oxide stones even though they're equal in grit will result in different finishes.
As you can see from the SAE paper, the Fuji engineers go to great lengths to address every problem. Their engines are jewels. I've worked on almost every brand and I really like what they did on the EJ engines. Only a couple small details I'd change.
3 and 4 are the closest ones to the flywheel Dan.
So how did the barbed fitting come apart? The Aeroquip hose is burst rated to 400 psi, pretty sure the oil filter and cooler will explode way before that happens. Even with the clamp removed, I can put both feet against the bench and pull as hard as I can- over 200 lbs. I am guessing, you can't pull the line off the fitting. This is the last thing I'm worried about. Done right, they simply don't fail.
valleys no since the ring never touches those
I know I asked the question, but after some thought I'm hard pressed to describe the mode shape.
Related or not, it sure would be easy to install a Centaflex while you have the Massive Marcotte off the engine. It should lower the resonant RPM, and hammer with less gusto when passing through it.
It was an Aeroquip socketless hose and it just blew the hose off.
Unless there is enough valley depth the rings push the oil ahead of the rings instead of displacing it in the valleys. Oil getting pushed to the combustion chamber = oil consumption. As you know well here.
So how did the barbed fitting come apart? The Aeroquip hose is burst rated to 400 psi, pretty sure the oil filter and cooler will explode way before that happens. Even with the clamp removed, I can put both feet against the bench and pull as hard as I can- over 200 lbs. I am guessing, you can't pull the line off the fitting. This is the last thing I'm worried about. Done right, they simply don't fail.
Beware that there is a significant difference in the way a hose behaves when pressurized internally vs. being pulled. When a hose is pressurized it will expand in diameter, but when it is pulled the hose can actually tighten in diameter slightly. This can mislead you into thinking that it's harder for a hose to come off a barbed fitting than it really is in service.
Every modern car that I've seen has gone away from barbed fittings for high pressure fuel plumbing, and I think for good reason: Reliability. If you look at Formula 1 or Indycar engines, they are all AN plumbing.
Frankly, using barbed fitting for oil and fuel lines in an aircraft is risky, and just plain (plane?) silly! Also the use of brass fittings in a vibration environment is asking for trouble because brass will work harden and can eventually crack.
Skylor
I am not picturing what you are saying here.
The oil starts out below the piston and the rings are designed to keep it there. The more texture there is in the cylinder wall, the more oil will stay there and not be scraped back down by the rings, like using a rubber squeegee on brushed concrete vs a piece of glass.
This engine has 5 main bearings so I'm thinking some nasty harmonic would have a hard time reaching #3 rod journal but I don't really know anything about that stuff.
First I'm trying more flywheel mass..... If this doesn't make a big difference, I have taken some measurements and will try fitting a Centaflex next winter...
The rings require a measured amount of oil volume from the cylinder walls. The hone finish does the metering. If the valleys are too deep, the engine will use oil. Too shallow, the engine will use oil. Some ring manufacturers give very specific crosshatch angles, like 40-45 degrees which you should easily be able to get precisely on a CK-10. Too much crosshatch angle, oil consumption will be higher over time. Too little, oil consumption will be higher over time.
You did take the engine apart due to high oil consumption right?
One more thing, the contact area of the valves looks a little wide, and one one of your intake valves looks like the lapped area sits a little too high. Since you're using a Neway seat cutting system there is a guy on ebay that sells blades with all three angles ground into a single set, check it out: http://www.ebay.com/itm/3-angle-val..._Automotive_Tools&vxp=mtr&hash=item19d8084fb3 They work really well.
The "nasty harmonic" is reaching all your rod bearings every time you pass through that big F1 at what, 1450? The crank twists one way, then the other. The question is the amplitude of the angular oscillation at each inertia (the rod throws) along the length of the crank. A "map" of the angular displacements is called a mode shape.
This particular set of mode shapes is an illustration taken from DenHartog's "Mechanical Vibrations". It's for a diesel motor generator set. In the context of this discussion, the key concept is the angular displacements expressed as a decimal percentage of the largest angular displacement, and their distribution along the train of inertias and stiffnesses.
What I've suggested in your case is the possibility of large angular oscillation at #'s 3 and 4. It's not hard to predict the mode shapes (first edition of the above was published in 1934) if you know the inertias and connecting stiffnesses. Unfortunately, we don't in this case so I can't prove anything. I can't even guess, given the long train (belted overhead cams, the vac pump, etc) and the oddities of the urethane bushings as a soft element (they are awful).
Why half-fix when you can do both now? I'm just sayin'
If the gearboxes were breaking up or the bushings were lasting only 50 hours, I'd have to instrument the thing somehow to learn what is actually happening.
The reason for trying the increased inertia first is about the scientific method [
But with crude calculation from this it would seem that peak loading may easily be in excess of 1000 lb./ft and possibly far above that figure.
You have an unexplained rod bearing breakup, apparently unique to your setup.
I sincerely beg your pardon, but the scientific exploration of torsional issues was 95% done by WWII. You're a smart guy. Quit pumping the pooch and apply the known.
You just had the engine apart. You know about bifillar suspension to determine inertia, but you didn't do it.
With the lump inertia of the crank/flywheel assembly and the lump inertia of the prop/gearbox you could select an off-the-shelf NBR elastomeric torsional coupler with a torsional stiffness and torque capacity listed right next to the part number. You know, an engineered component.
Yep, thus the noise and shaking airplane. What do you figure a 1000 lb-ft torque oscillation applied to the crank at 45 hertz is doing to the rod throws?
I must say that is has been a very "spirited" post. All of you have very valid points. I would say from my experience that there is a couple of over looked areas.
First. Whose piston is it? Some manufacture can't cut a flat ring grove to save their live.
Sec. What is the oil ring tension? With aftermarket rings there are many choices.
Third. What is the actual blow by in CFM? There is a meter to measure this.
As far as what "grit" to use. Forget it. Use RPK,RVK,PK and VO numbers based on what the ring is made of. Just remember. Crome won't get you home. There are many other materials that are much better. Total Seal has a ring I use in extreme application that I swore by. C23 is by far the most robust and has a easier time of sealing. Again the numbers are crucial. It has a carbon nitride face. C72 is a titanium nitride face but is a little fussier. How about cross hatch angle. A steep angle always produces more blow by. To flat and the ring will ?skate? across the bore and wear quickly. I found that 26-29 degrees is about right for any application. This is proven with a blow by meter and time in service.
As far as a CK-10, good machine in its day but has issues. Main one is ?loabing? cylinder. Just because the mike says its straight and round doesn?t mean its straight and round. How about a torque plate when you honed it?
I could go on and on with honing. With many years of perfecting it in NHRA Pro Stock competition I can say it is never perfect. I you have any questions and I have not pissed anyone off with my first post I am sure I could help you out.
By the way, cool forum. Looking into a RV8 TD and can?t wait to get my hands on one.